Karst groundwater and pollutant migration channel detection method and system

Abstract

The invention relates to the technical field of karst groundwater detection, and discloses a karst groundwater and pollutant migration channel detection method and system.The karst groundwater and pollutant migration channel detection method comprises the steps that a karst groundwater and pollutant migration range is preliminarily predicted to obtain a preliminary migration range; acquiring groundwater information and karst information of the preliminary migration range, acquiring pollution information, preprocessing the groundwater information, the karst information and the pollutant information with time sequences, inputting the preprocessed information as a prediction model, outputting the migration channel map as the prediction model, and constructing a karst groundwater and pollutant migration channel prediction model; the method comprises the following steps: constructing a karst groundwater and pollutant migration channel prediction model, obtaining groundwater information, karst information and pollutant information, and predicting a migration channel based on the karst groundwater and pollutant migration channel prediction model. And timely and effective recent emergency treatment and long-term pollution treatment are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of karst groundwater detection, and in particular to a method and system for detecting karst groundwater and pollutant migration channels. Background Art

[0002] Karst groundwater systems are widely distributed around the world due to their unique geological structure and hydrological characteristics. However, due to the development of fissures and caves in karst areas, groundwater flows rapidly and pollutants can spread rapidly through these channels, leading to increasingly serious groundwater pollution problems. Traditional groundwater pollution assessment and detection methods face many challenges in their application in karst areas.

[0003] Traditional methods for assessing and detecting groundwater pollution include geological surveys and mapping, geological radar, seismic exploration, hydrogeochemical methods, groundwater demonstration tests, geophysical detection methods, remote sensing technology, numerical simulation, etc. The above methods have defects such as low temporal and spatial resolution and large environmental impact. In addition, karst groundwater is affected by rainfall and seasonal changes. Heavy rain or human activities can cause sudden changes in groundwater levels and flow rates. There are often skylights and collapses in karst areas. There is a direct connection between groundwater and surface water, and there is often a recharge and discharge relationship between underground rivers and surface rivers. The karst groundwater and pollutant migration channels obtained by traditional methods are static channels with low accuracy, which is not conducive to timely and effective short-term emergency treatment and long-term pollution control. Summary of the invention

[0004] The present invention aims to provide a method and system for detecting karst groundwater and pollutant migration channels. By constructing a karst groundwater and pollutant migration channel prediction model, the dynamic channels of karst groundwater and pollutants are obtained, so as to achieve timely and effective short-term emergency treatment and long-term pollution control.

[0005] In order to achieve the above object, the present invention adopts the following technical scheme:

[0006] Karst groundwater and pollutant migration channel detection methods, including:

[0007] The preliminary prediction step is to make a preliminary prediction of the migration range of karst groundwater and pollutants to obtain a preliminary migration range;

[0008] Karst groundwater information acquisition steps, obtaining groundwater information and karst information of the preliminary migration range;

[0009] A pollution information acquisition step, obtaining pollution information, wherein the pollution information includes pollutant characteristic indicators, pollutant sources and continuous emission conditions;

[0010] The prediction model creation step is to pre-process the groundwater information, karst information and pollutant information with time series as the prediction model input, and use the migration channel map as the prediction model output to build the karst groundwater and pollutant migration channel prediction model;

[0011] The prediction and analysis step obtains groundwater information, karst information and pollutant information, and predicts the migration channel based on the karst groundwater and pollutant migration channel prediction model.

[0012] The principle and advantages of this scheme are: in actual application, according to field surveys, data access and other work, karst groundwater information and pollution information can be obtained, and information on the main migration channels of karst groundwater and pollutants can also be obtained. However, due to the unique geological structure and hydrological characteristics of the karst groundwater system, the dynamic changes of karst groundwater, and the complex water quality reaction of pollutants during migration, coupled with the influence of fissures and caves, groundwater and pollutants migrate through multiple paths such as fissures, caves, and pipelines, making it difficult to accurately obtain the migration channels of groundwater and pollutants through traditional means. By constructing a prediction model for the migration channel of karst groundwater and pollutants, the detection position of the migration channel of groundwater and pollutants can be quickly output, and it is easy to reflect the temporal and dynamic characteristics of the migration channel, so as to provide geological basis and basic support for emergency disposal of environmental pollution incidents, emergency treatment and post-management of contaminated karst springs, etc.

[0013] Preferably, as an improvement, the preliminary prediction step includes:

[0014] a water level measurement sub-step, measuring the groundwater level of the monitoring well;

[0015] The hydraulic gradient calculation sub-step calculates the hydraulic gradient based on the water level difference of the monitoring well. The hydraulic gradient calculation model is:

[0016]

[0017] Among them, m i is the hydraulic gradient, Δh is the water level difference, and l is the dividing horizontal distance;

[0018] The velocity calculation sub-step obtains the groundwater velocity based on Darcy's law. The groundwater velocity calculation model is:

[0019] v im =k*m i

[0020] Among them, v im is the groundwater flow rate and k is the permeability coefficient.

[0021] Technical effect: By obtaining the groundwater flow velocity, it is convenient to obtain the preliminary migration range.

[0022] Preferably, as an improvement, the initial migration range includes the set of all groundwater runoff zones connected to the pollution source {L 1 , L 2 , …, L i},in:

[0023]

[0024] is the average groundwater velocity in the groundwater runoff zone, t is the expected treatment time, L i is the length range of groundwater runoff, L d It is the length to the end of groundwater.

[0025] Technical effect: Defining the initial migration scope will make it easier to obtain data and reduce manpower and material expenditures.

[0026] Preferably, as an improvement, the prediction model is a CNN convolutional neural network model.

[0027] Technical effect: It can realize the feature extraction of complex groundwater systems and improve the accuracy of prediction results.

[0028] Preferably, as an improvement, the prediction model creation step includes:

[0029] The data preprocessing sub-step cleans and standardizes the data and creates feature engineering;

[0030] The data conversion sub-step converts the pre-processed data into images;

[0031] The model construction sub-steps accept image data in the input layer, extract features in the convolution layer, use the pooling layer to reduce the dimension of the feature map, use the ReLU activation function, map the features extracted by the convolution layer to the fully connected layer for classification, and the output layer outputs the predicted migration channel map.

[0032] Technical effect: By creating a predictive model, it is easy to quickly obtain the migration channel.

[0033] Preferably, as an improvement, the feature engineering includes groundwater level change rate and pollutant concentration gradient.

[0034] Technical effect: The groundwater level and pollutant concentration are dynamically changing characteristics. Through feature extraction, it is easy to regularize the irregular data.

[0035] Preferably, as an improvement, the data visualization includes: converting the geological profile of the preliminary migration range into an image format, and converting the groundwater information and pollution information into a two-dimensional image.

[0036] Technical effect: Image processing makes it easier to improve the processing efficiency of the CNN convolutional neural network model.

[0037] Preferably, as an improvement, a visualization step is also included, using geographic information system software to display the predicted pollutant migration channels in the form of a map.

[0038] Technical effect: It is easy to improve the visualization of results.

[0039] Preferably, as an improvement, the method further includes a step of analyzing the distribution patterns of pollutant migration, and based on the prediction results of karst groundwater and pollutant migration channels, the spatial distribution and temporal variation characteristics of pollutant migration are analyzed and obtained.

[0040] Technical effect: Provide basis and basic support for the selection of emergency disposal projects for springs, the selection of long-term treatment plans, and the layout of subsequent monitoring projects.

[0041] It also includes a karst groundwater and pollutant migration channel detection system, which uses the karst groundwater and pollutant migration channel detection method. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 Schematic diagram of the flow chart of the method for detecting karst groundwater and pollutant migration channels. DETAILED DESCRIPTION

[0043] The following is further described in detail through specific implementation methods:

[0044] The embodiment is basically as shown in the attached Figure 1 As shown:

[0045] Karst groundwater and pollutant migration channel detection methods, including:

[0046] The preliminary prediction step is to make a preliminary prediction of the karst groundwater and the migration range of pollutants to obtain a preliminary migration range; obtaining the preliminary migration range is convenient for reducing the difficulty of data acquisition and reducing human and material expenditures. The preliminary migration range includes the set of all groundwater runoff zones connected to the pollution source {L 1 , L 2 , …, L i},in:

[0047]

[0048] is the average groundwater velocity in the groundwater runoff zone, t is the expected treatment time, L i is the length range of groundwater runoff, the length range is the plane range, L d It is the length at the end of the groundwater, that is, the plane length from the pollution source to the end of the groundwater.

[0049] The preliminary prediction step includes a water level measurement sub-step, a hydraulic gradient calculation sub-step and a flow rate calculation sub-step. The water level measurement sub-step measures the groundwater level of the monitoring well; the hydraulic gradient calculation sub-step calculates the hydraulic gradient based on the water level difference of the monitoring well. The hydraulic gradient calculation model is:

[0050]

[0051] Among them, m i is the hydraulic gradient, Δh is the water level difference, and l is the dividing horizontal distance;

[0052] The velocity calculation sub-step obtains the groundwater velocity based on Darcy's law. The groundwater velocity calculation model is:

[0053] v im =k*m i

[0054] Among them, v im is the groundwater flow rate and k is the permeability coefficient.

[0055] The step of obtaining karst groundwater information is to obtain groundwater information and karst information of the preliminary migration range. In this embodiment, it includes but is not limited to conducting geophysical exploration, drilling, tracer test, water injection test, water and soil sampling and analysis and other exploration work to obtain groundwater information and karst information. The groundwater information includes groundwater runoff zone and runoff zone characteristics. The groundwater runoff zone is a planar position. The runoff zone characteristics include groundwater flow velocity, hydraulic gradient, and groundwater flow field; the karst information includes the development and distribution of karst fissures. The karst area has developed fissures and a complex fissure network, which leads to uneven distribution of groundwater and pollutants in the fissures. The existence of caves makes the groundwater flow path complex and changeable, and pollutants are easily gathered or dispersed in the caves. Karst information is used to facilitate accurate extraction of the migration law of pollutants.

[0056] The step of obtaining pollution information includes obtaining pollution information, wherein the pollution information includes pollutant characteristic indicators (e.g., when an oily wastewater overflow occurs, the pollutant characteristic indicators include the concentrations of petroleum hydrocarbons (C10-C40), benzo[a]pyrene, dibenzo[a,h]anthracene, pentachlorophenol, etc.), pollutant sources, and continuous emission conditions.

[0057] The karst groundwater system has a unique geological structure and hydrological characteristics, and the migration of groundwater and pollutants is complex. For example, cracks are well developed in karst areas, and the crack network is complex, resulting in uneven distribution of groundwater and pollutants in the cracks. The high permeability of groundwater leads to a fast migration speed of pollutants in groundwater. The existence of caves makes the groundwater flow path complex and changeable, and pollutants are easily gathered or dispersed in caves. The karst groundwater has high water velocity and flow in the main flow channel (karst pipe), and the migration speed of pollutants in the pipe is fast. The groundwater level changes caused by external factors such as rainfall will be rapidly transmitted to the entire groundwater system, and the migration of pollutants will respond instantaneously. The seasonal changes in rainfall will affect the groundwater recharge and groundwater level, and then affect the migration of pollutants. The seasonal changes in evaporation will affect the balance of groundwater recharge and consumption, and affect the migration of pollutants. Heavy rain events will cause the groundwater level to rise rapidly, and pollutants will migrate rapidly through cracks and caves. During the migration of pollutants in groundwater, there are chemical reactions with minerals in the groundwater, which change the form and toxicity of pollutants. During the migration of pollutants in groundwater, pollutants will be adsorbed by minerals on the walls of cracks and caves. When environmental conditions change, pollutants adsorbed on the walls of cracks and caves may be re-desorbed and continue to migrate. Therefore, how to extract features from complex karst groundwater systems to predict migration pathways has become a difficult problem.

[0058] In this embodiment, the prediction model creation step uses the time-series groundwater information, karst information and pollutant information as the prediction model input after preprocessing, and uses the migration channel map as the prediction model output to build a karst groundwater and pollutant migration channel prediction model; the preferred prediction model is the CNN convolutional neural network model. The powerful feature extraction capability of the CNN convolutional neural network model can realize the feature extraction and prediction of complex karst groundwater systems.

[0059] The prediction model creation step includes a data preprocessing sub-step, a data conversion sub-step, and a model framework sub-step. The data preprocessing sub-step cleans and standardizes the data and creates feature engineering; specifically, a large amount of historical data is obtained, and the historical data is cleaned to deal with missing values, outliers, and duplicate values; the cleaned data is converted to the same scale, that is, standardized, to improve the availability of the data. The feature engineering includes the groundwater level change rate and the pollutant concentration gradient. The groundwater level change rate and the pollutant concentration gradient are important parameters for evaluating groundwater flow and pollutant migration. The groundwater level change rate calculation model is:

[0060]

[0061] Among them, h t+1 and h t are the groundwater levels at time t and t+1 respectively, and Δt is the time interval.

[0062] The calculation model of pollutant concentration gradient is:

[0063]

[0064] Among them, c 2 、c 1 are the pollutant concentrations at monitoring point 2 and monitoring point 1, respectively, and Δx is the distance between the two monitoring points.

[0065] The data conversion sub-step visualizes the pre-processed data; the data visualization includes: organizing the karst information of the preliminary migration range into a geological profile, converting the geological profile into an image format, and converting the groundwater information and pollution information into a two-dimensional image. For example, the groundwater level data is converted into a two-dimensional image, and each pixel represents the groundwater level at a location, and the pollutant concentration data is converted into a two-dimensional image, and each pixel represents the pollutant concentration at a location. After data conversion, data enhancement is also required, such as rotating and flipping the image data to increase the diversity of the data; adding noise to the image to improve the robustness of the model.

[0066] The model construction sub-steps accept image data in the input layer, extract features in the convolution layer, use the pooling layer to reduce the dimension of the feature map to extract the main features, use the ReLU activation function to map the features extracted by the convolution layer to the fully connected layer for classification, and the output layer outputs the predicted migration channel map. After the model is trained and optimized, it can be used for predictive analysis.

[0067] It also includes a visualization step, using geographic information system software to display the predicted pollutant migration channel in the form of a map to improve the visualization of the results. It also includes a pollutant migration distribution law analysis step, based on the karst groundwater and pollutant migration channel prediction results, to analyze and obtain the spatial distribution and temporal variation characteristics of pollutant migration. For example, the spatial distribution of the predicted pollutant migration channel is analyzed to determine the migration range of pollutants, the spatiotemporal variation of the predicted pollutant migration channel is analyzed, and the migration path and speed of pollutants are understood.

[0068] It also includes a karst groundwater and pollutant migration channel detection system, which uses the karst groundwater and pollutant migration channel detection method.

[0069] The above is only an embodiment of the present invention, and the common knowledge such as the known specific technical solutions and / or characteristics in the solution is not described in detail here. It should be pointed out that for those skilled in the art, without departing from the technical solution of the present invention, several modifications and improvements can be made, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. Karst groundwater and pollutant migration channel detection method, characterized by ,include: The preliminary prediction step is to make a preliminary prediction of the migration range of karst groundwater and pollutants to obtain a preliminary migration range; Karst groundwater information acquisition steps, obtaining groundwater information and karst information of the preliminary migration range; A pollution information acquisition step, obtaining pollution information, wherein the pollution information includes pollutant characteristic indicators, pollutant sources and continuous emission conditions; The prediction model creation step is to pre-process the groundwater information, karst information and pollutant information with time series as the prediction model input, and use the migration channel map as the prediction model output to build the karst groundwater and pollutant migration channel prediction model; The prediction and analysis step obtains groundwater information, karst information and pollutant information, and predicts the migration channel based on the karst groundwater and pollutant migration channel prediction model.

2. The method for detecting karst groundwater and pollutant migration channels according to claim 1, characterized in that , the preliminary prediction step includes: a water level measurement sub-step, measuring the groundwater level of the monitoring well; The hydraulic gradient calculation sub-step calculates the hydraulic gradient based on the water level difference of the monitoring well. The hydraulic gradient calculation model is: Among them, m i is the hydraulic gradient, Δh is the water level difference, and l is the dividing horizontal distance; The velocity calculation sub-step obtains the groundwater velocity based on Darcy's law. The groundwater velocity calculation model is: in im =k*m i Among them, v im is the groundwater flow rate and k is the permeability coefficient.

3. The method for detecting karst groundwater and pollutant migration channels according to claim 2 is characterized in that , the initial migration range includes the set of all groundwater runoff zones connected to the pollution source {L1, L2, ..., L i },in: is the average groundwater velocity in the groundwater runoff zone, t is the expected treatment time, L i is the length range of groundwater runoff, L d It is the length to the end of groundwater.

4. The method for detecting karst groundwater and pollutant migration channels according to claim 1, characterized in that: The prediction model is a CNN convolutional neural network model.

5. The method for detecting karst groundwater and pollutant migration channels according to claim 4 is characterized in that ,The prediction model creation step includes: The data preprocessing sub-step cleans and standardizes the data and creates feature engineering; The data conversion sub-step converts the pre-processed data into images; The model construction sub-steps accept image data in the input layer, extract features in the convolution layer, use the pooling layer to reduce the dimension of the feature map, use the ReLU activation function, map the features extracted by the convolution layer to the fully connected layer for classification, and the output layer outputs the predicted migration channel map.

6. The method for detecting karst groundwater and pollutant migration channels according to claim 5, characterized in that: The feature engineering includes groundwater level change rate and pollutant concentration gradient.

7. The method for detecting karst groundwater and pollutant migration channels according to claim 5, characterized in that: The data visualization includes: converting the geological profile of the preliminary migration range into an image format, and converting the groundwater information and pollution information into a two-dimensional image.

8. The method for detecting karst groundwater and pollutant migration channels according to claim 7, characterized in that: It also includes a visualization step, using geographic information system software to display the predicted pollutant migration pathways in the form of a map.

9. The method for detecting karst groundwater and pollutant migration channels according to claim 8, characterized in that: It also includes the step of analyzing the migration and distribution patterns of pollutants. Based on the prediction results of karst groundwater and pollutant migration channels, the spatial distribution and temporal variation characteristics of pollutant migration are analyzed and obtained.

10. Karst groundwater and pollutant migration channel detection system, characterized by: Any karst groundwater and pollutant migration channel detection method described in claims 1-9 is used.

Images